In addition to stated behavioral time-series data, an extensive set of personal sociodemographic, profile (age, gender, ethnicity, educational and marital status, employment status and occupational environment…), as well as health state, risk factors and
habits, lifestyle, neighbourhood and quality of life assessment, and other relevant
behavioural data are manually input/submitted on each citizen participating in the study
via online forms, composed from adapted relevant survey/assessment instruments for
each specific field, like Framingham, EuroQoL-5, IPAQ-SF. These data are geolocalized to the residence location of each responding citizen for the purposes of
analytics of collective/community well-being, and are collected from a greater number
of recruited respondents, but just in rare cases in more than one iteration over time due
to the high number and scope of covered variables, and therefore suitable for a broad
but mainly static “snapshot” assessment of current well-being state rather than for
behaviour change model and analytics. Incorporating both these static and IoT-sensed
temporal data into a fully comprehensive predictive well-being model is an ongoing
task in progress throughout the end of the Project, with results to be presented in other
upcoming publications.
4 Derivation of Domain/Dimension Indices
4.1 Physical Activity/Exercise
Physical activity in our first approach stated above in Sect. 2 can be discretized using
several common baseline categorizations related or derived from the above mentioned
relevant institutional/governmental and professional expert guidelines for the urban
population groups. The example approach taken in the recent health survey of England
from 2016 [12] compared well-being and mental health of adults in different sociodemographically stratified population groups by physical activity, among others. The
activity level categories used in the analysis were the following:
• Meets aerobic guidelines: At least 150 min moderately intensive physical activity
or 75 min vigorous activity per week or an equivalent combination of these
• Asserted activity: 60 to 149 min moderate activity or 30–74 min vigorous activity
per week or an equivalent combination of these
• Low activity: 30 to 59 min moderate activity or 15 to 29 min vigorous activity per
week or an equivalent combination of these
• Inactive: Less than 30 min moderate activity or less than 15 min vigorous activity
per week or an equivalent combination of these,
and the corresponding linear scaled scoring function denotes “Meets aerobic
guidelines” with a score of 4, “Certain activity” - 3, “Low activity” - 2, and “Inactive”
with 1. This baseline scoring scale, besides sufficient granularity and robustness
exhibited in referenced comprehensive studies, is also convenient for
• mapping to the defined activity level categories used as input parameters for the
consensus models for prediction of risk of Type 2 Diabetes (T2D) and asthma onset
and exacerbation, developed for the PULSE project [13, 17]
160
V. Urošević et al.
habits, lifestyle, neighbourhood and quality of life assessment, and other relevant
behavioural data are manually input/submitted on each citizen participating in the study
via online forms, composed from adapted relevant survey/assessment instruments for
each specific field, like Framingham, EuroQoL-5, IPAQ-SF. These data are geolocalized to the residence location of each responding citizen for the purposes of
analytics of collective/community well-being, and are collected from a greater number
of recruited respondents, but just in rare cases in more than one iteration over time due
to the high number and scope of covered variables, and therefore suitable for a broad
but mainly static “snapshot” assessment of current well-being state rather than for
behaviour change model and analytics. Incorporating both these static and IoT-sensed
temporal data into a fully comprehensive predictive well-being model is an ongoing
task in progress throughout the end of the Project, with results to be presented in other
upcoming publications.
4 Derivation of Domain/Dimension Indices
4.1 Physical Activity/Exercise
Physical activity in our first approach stated above in Sect. 2 can be discretized using
several common baseline categorizations related or derived from the above mentioned
relevant institutional/governmental and professional expert guidelines for the urban
population groups. The example approach taken in the recent health survey of England
from 2016 [12] compared well-being and mental health of adults in different sociodemographically stratified population groups by physical activity, among others. The
activity level categories used in the analysis were the following:
• Meets aerobic guidelines: At least 150 min moderately intensive physical activity
or 75 min vigorous activity per week or an equivalent combination of these
• Asserted activity: 60 to 149 min moderate activity or 30–74 min vigorous activity
per week or an equivalent combination of these
• Low activity: 30 to 59 min moderate activity or 15 to 29 min vigorous activity per
week or an equivalent combination of these
• Inactive: Less than 30 min moderate activity or less than 15 min vigorous activity
per week or an equivalent combination of these,
and the corresponding linear scaled scoring function denotes “Meets aerobic
guidelines” with a score of 4, “Certain activity” - 3, “Low activity” - 2, and “Inactive”
with 1. This baseline scoring scale, besides sufficient granularity and robustness
exhibited in referenced comprehensive studies, is also convenient for
• mapping to the defined activity level categories used as input parameters for the
consensus models for prediction of risk of Type 2 Diabetes (T2D) and asthma onset
and exacerbation, developed for the PULSE project [13, 17]
160
V. Urošević et al.